
\begin{table}[htb]
\caption{Robustness check: Table F3 with alternative operationalisation of role conceptions}
\begin{center}
\begin{tabular}{l c c c}
\hline
 & Experiment 1 & Experiment 2 & Experiment 3 \\
\hline
Citizens                                               & $-0.21$           & $1.84^{*}$        & $-0.73$           \\
                                                       & $ [-0.70;  0.33]$ & $ [ 1.30;  2.39]$ & $ [-1.63;  0.10]$ \\
Benefits                                               &                   &                   & $4.12^{*}$        \\
                                                       &                   &                   & $ [ 3.34;  4.87]$ \\
Citizens $\times$ Benefits                             &                   &                   & $0.67$            \\
                                                       &                   &                   & $ [-0.44;  1.71]$ \\
Agree with D3                                          & $-1.39^{*}$       & $-0.80^{*}$       & $-0.27$           \\
                                                       & $ [-1.91; -0.88]$ & $ [-1.38; -0.19]$ & $ [-1.10;  0.58]$ \\
Str. agree with D3                                     & $-2.89^{*}$       & $-1.99^{*}$       & $-1.29^{*}$       \\
                                                       & $ [-3.45; -2.29]$ & $ [-2.58; -1.28]$ & $ [-2.22; -0.36]$ \\
Population $\times$ Agree with D3                      & $1.75^{*}$        & $0.36$            & $-1.03$           \\
                                                       & $ [ 1.02;  2.42]$ & $ [-0.38;  1.11]$ & $ [-2.14;  0.04]$ \\
Population $\times$ Str. agree with D3                 & $3.45^{*}$        & $1.17^{*}$        & $-0.43$           \\
                                                       & $ [ 2.65;  4.20]$ & $ [ 0.17;  2.04]$ & $ [-1.71;  0.85]$ \\
Benefits $\times$ Agree with D3                        &                   &                   & $0.02$            \\
                                                       &                   &                   & $ [-0.95;  1.01]$ \\
Benefits $\times$ Str. agree with D3                   &                   &                   & $0.76$            \\
                                                       &                   &                   & $ [-0.41;  1.94]$ \\
Citizens $\times$ Benefits $\times$ Agree with D3      &                   &                   & $0.98$            \\
                                                       &                   &                   & $ [-0.33;  2.28]$ \\
Citizens $\times$ Benefits $\times$ Str. agree with D3 &                   &                   & $0.49$            \\
                                                       &                   &                   & $ [-1.19;  2.15]$ \\
Intercept                                              & $6.58^{*}$        & $4.66^{*}$        & $4.21^{*}$        \\
                                                       & $ [ 6.19;  6.99]$ & $ [ 4.19;  5.12]$ & $ [ 3.61;  4.83]$ \\
\hline
N                                                      & $863$             & $863$             & $863$             \\
\hline
\multicolumn{4}{l}{\scriptsize{\parbox{0.95\linewidth}{$^*$ Null hypothesis value outside the 95\% confidence interval. Entries are unstandardized coefficients from a linear mixed-effect regression model. Bootstrapped confidence intervals in parentheses. Countries function as groups. Support for PTAs is captured on an 11-point scale. Item D3 reads `The influence of economic elites on trade negotiations is too big.'}}}
\end{tabular}
\label{tab:tableH7}
\end{center}
\end{table}
